Geometric complement heterogeneous information and random forest for predicting lncRNA-disease associations

Area of Science:

• Genomics
• Bioinformatics
• Computational Biology

Background:

• Aberrant expression of long non-coding RNAs (lncRNAs) is linked to various human diseases.
• Accurate identification of disease-related lncRNAs is crucial for understanding molecular mechanisms and developing effective treatments.
• Existing lncRNA-disease association prediction models face challenges in identifying unknown associations.

Purpose of the Study:

• To develop a novel computational model for predicting lncRNA-disease associations.
• To improve the accuracy and efficiency of identifying unknown lncRNA-disease relationships.
• To provide a tool for exploring potential therapeutic targets related to lncRNAs.

Main Methods:

• Integration of lncRNA-miRNA interactions and miRNA-disease associations using geometric complement heterogeneous information.
• Feature engineering for lncRNAs and diseases based on similarity coefficients.
• Dimensionality reduction using an autoencoder for feature representation learning.
• Classification using a random forest model trained on fused low-dimensional features.

Main Results:

• The proposed model achieved an Area Under the Receiver Operating Characteristic Curve (AUC) of 0.9897 and an Area Under the Precision-Recall Curve (AUPR) of 0.7040 in five-fold cross-validation.
• Performance surpasses several state-of-the-art lncRNA-disease association prediction models.
• Case studies on colon and stomach cancer validated the model's capability in predicting disease-related lncRNAs.

Conclusions:

• The developed computational model effectively predicts lncRNA-disease associations.
• The approach offers a promising strategy for discovering novel disease-related lncRNAs.
• This work contributes to advancing the understanding of lncRNA functions in human diseases and potential therapeutic interventions.